Benedict Evans is an independent analyst and former partner at Andreessen Horowitz, where he spent years as their in-house “thinker” tracking the most important technology trends. For the past six years, he’s been publishing deeply researched presentations on where tech is heading, most recently foc
Recognize that we are in the '1997 of AI', where there is great uncertainty about what will come next. As a PM, you must build flexible strategies that adapt to multiple possible futures instead of one-directional bets.
We're in 1997 for AI—early, exciting, and deeply uncertain about what comes next
Distribution is becoming the ultimate moat in AI-driven software. As base models democratize and the cost of creating AI features decreases, the ability to bring products to users will be more valuable than the technical capability to build them.
Distribution is becoming the ultimate moat as software gets easier to build
Reframe the question about AI and jobs: don't ask 'what percentage of this job can AI do', ask 'is this a task or a job'. Tasks can be automated but jobs are more resilient because they involve human context, judgment, and relationships.
Why the right question about your job isn't 'What percent can AI do?' but 'Is this a task or a job?'
Observe the explosion in demand for consulting and professional services at AI companies. This signals that value is accumulating in the ability to apply AI to specific problems, not in building base models. Rebuild your product mindset around AI application versus AI construction.
The surprising boom in consulting and professional services at AI companies
Prepare yourself for economic and social changes without assuming they will be catastrophic. Focus on building adaptability skills in yourself and your team: learning new tools quickly, understanding context in new domains, working effectively with technology that changes every 18 months.
Why things will probably be okay—and what you need to do to prepare